vector multiplication .* vs *
399 ビュー (過去 30 日間)
I have these two vectors,
u= [ 1 2 -1 2 1]
v=[ -1 0 2 0 1]
why does v*u give me an error, whilst u*v' give me a -2, which is the result of a scalar multiplication?
u and v have the same size, so shouldn't v*u go through?
Thanks in advance.
Walter Roberson 2017 年 11 月 19 日
v*u is algebraic matrix multiplication, for which the rule is that for an (M x N) * (P x Q) operation, that N must equal P and the output size is M x Q -- so (M x N) * (N x Q) giving M x Q. You have (1 x 5) * (1 x 5) which violates that rule. When you take u * v' then you have (1 x 5) * (5 x 1) giving 1 x 1.
v.*u would go element by element, result(K) = v(K) * u(K) which would be 1 x 5 result.
その他の回答 (2 件)
John Keevil 2022 年 2 月 18 日
The simple answer to the question is:-
u= [ 1 2 -1 2 1] and v=[ -1 0 2 0 1] are both row vectors. You cannot matrix multiply them because the number of rows and columns are not compatible for matrix multiplication. To multiply u*v by matrix multiplication requires the number of rows of u to equal the number of columns of v, which it does not since u has one row and v has 5 columns. To correct that, you have to transpose v, then it has one column. v' is the transpose of v. This why u*v' works.